Pre-Amp Clipping

[charczar27]

Hey all, first post here!

I'm working on designing a 50 Watt Push/Pull amp. This is my first try at amp building, and understand that I chose a pretty rough first build and probably 'in over my head'. However, I've dumped some time into this already so I'm pushing through until it's completed lol, I'm pretty confident I can figure it out.

I've been relying heavily on running simulations of my pre-amp/phase inverter through Multi-sim, and this is where most of my questions have originated. I know that from an engineering perspective guitar amps are poorly designed, and as a student I've been scratching my head QUITE a bit and summed it up as 'more art than science'. What's been confusing me is the point at which clipping occurs. I understand this is due to tube characteristics as well as the bias/'Q-point' of the said tube. The reason why it's been dragging me around in circles for weeks now is because I'm wanting to design an amp that has a lot of headroom in one circuit and have a separate circuit for overdrive. (I also understand that Overdrive is just a clipped signal) Separating the two circuits (switchable) is no biggie, my issue is due to the resulting sine wave from the output. In my head a pure sine wave (https://4.bp.blogspot.com/_6t_ZmJSkbL4/TJLFDrqFhYI/AAAAAAAACbU/BZgLx7HLP9U/w1200-h630-p-k-no-nu/Sinewave_horizontalview.jpg) is clean and has been what I've been aiming for. However, in most simulations I've been running I'm getting a severely clipped signal after the second tube (https://upload.wikimedia.org/wikipedia/commons/thumb/9/96/Voltage_Clipping.svg/1200px-Voltage_Clipping.svg.png). I know that the clipping is a result of the tube bias point, as well as the input signal; I also understand that the input signal is sweeping above the 0 Volt grid point in which the plate voltage is held constant until the input signal swings below the 0 grid Voltage point again.

I've ran multiple simulations of different bias points of 12AX7 and could obtain a clean sine wave at VERY low volume settings. The only operable design I've made that had a clean wave at the output for nearly all volume levels has been with 2 12AU7's, but I figure that this would be a bad design as it wouldn't drive the output tubes hard enough. (I guess I should have mentioned that I had a Tonestack in between these two tubes)

I figure that I am overthinking alot of this, and clipping in the Pre-amp is SUPPOSED to happen, regardless if it is the clean or overdrive channel. I even ran a simulation on the Fender Princeton Reverb AA764 and the clipping after the second tube was much more severe than in ANY of my simulations. However, I can't help but to continuously think and redesign, think and redesign.


So a couple of the questions that I have are:

Is clipping supposed to occur early in the pre-amp circuit? (As in before the Phase-Inverter)

Will the resulting (clipped) signal effect output power calculations? (I've already figured out my output circuit)

Are there are resources I can look to that demonstrates how the clipped signals will sound?



I'm sure this has been asked before, but honestly there's just so much information in this forum that I decided to make my own post. Plus, I figure I could get a little help from all the guru's out there.

Any kind of information you can give me I will appreciate greatly. 

[Ritchie200]

Can you give us an exact schematic of what you have in mind - tubes, tone stack, values of components.  That way we can see what you are doing.  I know you are not set on any one setup right now.  But there may be some serious basic flaws in the design that need to be cleaned up before you can move forward to tweak.

Jim

[SILVERGUN]

This doesn't address your questions exactly but it is required reading on the topic.
Download the PDF in this link: http://www.valvewizard.co.uk/gainstage.html

[charczar27]

If I remember correctly, this was my initial design. I hope you can read it just fine, I'll redraw abetter one if seeing it is an issue.

[charczar27]

My most recent design is Multi-Sim. Let me tweak few things inside of it so you can see those values.

**found it

[shooter]

Once you follow the link in reply 2 and read twice;
I came here from an Electronics background, basically RF, If my Bosses seen the wave shapes I get in guitar amps, On my work scope, I woulda been fired looong ago.  what you're gonna find, It's very hard to computerize the AC component of guitar amps, DC, easy enough, 'sept it has quite an impact on AC, (see PDF).

You have a pretty good start, I didn't look at the details, but the architecture, looks 5E3-ish.  maybe take a look at that schematic and compare against your values, since the Fender is a proven design 

[PRR]

> getting a severely clipped signal after the second tube

So turn-down the first volume/gain control. MultiSim may be too stupid to do that, but most musicians will figure it out.

[charczar27]

Thanks guys. I suffer from a lot of brain stuff and overthinking to the point of insanity is one of those symptoms lol. As far as turning down the volume I figured that out, I ran the simulation with volume settings and I was overthinking too much about the point at which it started clipping...the simulation started at the 'pre-amp volume = 3 and master volume = 3' point. I'm putting way too much thought into this I guess.

I'm just really paranoid about designing the amp, building it and blowing the output transformer not even 10 minutes into the first jam because of some mis-calculation I did concerning the guitar signal.

However - hypothetically speaking; if you did create an amp that amplified an absolutely clean sine wave all the way through - would this sound REALLY bright and thin? Is it the clipping of the wave that gives the guitar amp that punch? Ya know, like that 'oomf' you get with an E? 

[shooter]

absolutely clean sine wave

4 of the last 5 amps, I designed for 80% clean from pre, 80% clean PA, the last amp was built as a lead/distortion amp, complete square-wave outta the PA at max.  Nothing I've built, well, came from the either, it came from stitching pieces that are tried n true. 

If your sim shows DC characteristics within tubes "operational range" you shouldn't make smoke, bad sound, maybe

[92Volts]

Besides turning down the volume (which prevents clipping in both simulations and real life), what input signal are you simulating with? Guitar levels vary but 200mv is a good guess. If you simulate with a 1v input, you'll see clipping 5x sooner than you'd expect when playing a guitar through it.

Also, "audio" taper pots let about 10% of the signal through at the 1/2 setting. So unless you've accounted for that in your simulation... if you've found that you need 3/10 amplitude from the pot to avoid clipping, you can physically turn it more than halfway before that happens.

Clipping will not fry your OPT.

On the other hand, I'm not sure a lack of clipping will sound that bad... it could reduce some of the sound people usually look for from tube amps. But a lot of that comes from the signal being amplified asymmetrically, which produces even-order harmonics. This can occur long before clipping, where one side of the wave totally stops getting larger in response to a larger input.

[charczar27]

ok cool.

I was logarithmic pots and a 500 mV (peak) input. I read p90's had about a 240 mV(Peak) so doubled that in the case of humbuckers.

But thank you all, I think I can begin working on the project again.

[shooter]

240 mV(Peak)

I've settled on 130Vac rms based on the average pickup samples, maybe 7,8? so you're close enough 

enjoy the journey!

[charczar27]

Thanks for all the support!

[jjasilli]

@charczar27:  You have asked a umber of intersting questions:  I suggest some reference books:
Dave Funk's Tube Amp Workbook
Inside Tube Amps, Dan Torres
Electric Guitar Amplifier Handboo, Jack Darr
TUT1, Kevin O'Connor (pretty technical)
The Tube Amp Book, Aspen Pittman


Addressing some of your points
Is clipping supposed to occur early in the pre-amp circuit? (As in before the Phase-Inverter)
In the early days clipping was not desirable.  Eventually it became a Holy Grail.  Then, Mesa Boogie popularized preamps with cascading gain stages, specifically to sculpt overdrive tone in the preamp.

Will the resulting (clipped) signal effect output power calculations? (I've already figured out my output circuit)
Preamp tubes are not considered power devices, but rather voltage amplifiers.  Preamp overdrive does not effect power calculations.  If the PI & power tubes are not themselves in overdrive, then they will give a "clean boost" to the signal they receive from the preamp.  If the preamp signal is "dirty", there will be a clean boost to the dirty signal, with no change in the power calculation of the power tubes.

If the power tubes are in overdrive, that will affect the power calculation.

Are there are resources I can look to that demonstrates how the clipped signals will sound?
See the above reference books.  I believe there are pertinent oscilloscope screen shots on Bill Machrone's Blues Junior website.  Also checkout the following websites:  Aiken Amps; Rob Robinette; the Valve Wizzard; hope I haven't left anyone out.


Also research the difference between soft vs. hard clipping. 

[ALBATROS1234]

i am pretty new to tubes but not to electronics guitar or tube amps. i am working on my 6th amp now. in my opinion you are most definitely overthinking and stressing for no reason. my method was to read and study for a few months shematics of amps i liked and i started to see similar component values in certain places so i gathered parts and got out the soldering gun. when things didnt exactly sound like i wanted i read and asked questions tweaked accordingly... repeat...tweak ...until satisfied.what i have noticed is #1 this is very forgiving. things dont have to be exact to get a good sound. like if your output transformer is this or that impedance as long as its in range it will work, nothing will blow up it will just change your tone. hell my 1st couple amps i didnt even match impedance on the output transformr i had a single ended one from a powered speaker and popped it in and even though it was with a 25l6 power tube and i was using a 6f6 it worked fine. i even used a regular step down 277vac to 24vac trannie as an output transformer and it sounded better than one i paid $30 for and works great in the amp even though i have no clue the impedance . now my mad scientist methods are not for everyone but i am having a ball and seeing results. clipping is what you want to see too. even on clean channel if you got a moderate volume going you will have some slight clipping at least .personally i think if there was a perfect sine wave on the o-scope with a guitr input it would sound dry and flavorless.  tube clipping is why we love tube amps. personally if you are that worried perhaps you should build an amp type  that you like the tone of (princeton or whatever)then mod it to suit your taste. glancing at your schematic some of the values look odd compared to the schematics i have seen. one trick to reduce gain/clipping in an amp is to remove those kathode bypass caps or at least the 1st triodes bypass cap. you didnt put those value on the schematic. look at schematic of amps you want to mimic to get ideas of component value. the guys that designed them knew what they were doing . changing their values a bit will change the tone change them too much and it will be deficient in some way. at least this is how i see it.

[HotBluePlates]

... I've been relying heavily on running simulations of my pre-amp/phase inverter through Multi-sim ... However, in most simulations I've been running I'm getting a severely clipped signal after the second tube
...
Is clipping supposed to occur early in the pre-amp circuit? (As in before the Phase-Inverter) ...

Are you perhaps designing the amp starting at the input jack and moving through to the output?  The traditional approach yielding the most success starts at the speaker/output section and works backwards towards the input jack.

When you do that, you think in terms of how much drive signal the output tubes will need for maximum clean output power (anything beyond that level will be output tube distortion).  That tells you have big a drive signal you need from the phase inverter, and after design (when the phase inverter gain and feedback are known) how much drive signal the phase inverter needs.

And so on backwards through each stage towards the input jack.

... 500 mV (peak) input.  ...

If you investigate old amp designs, you might find the amp will distort the output tubes with the Volume control between half to 2/3, and with a signal level at the input jack of maybe 20-50mV.  In other words, crank the Volume and even the weakest pickups distort the output tubes.

Now if your goal is preamp tube distortion, add a master volume before the output tubes and add some more preamp gain.  Then it's up to you to decide "how much gain" and "where" and whether each stage adds a little distortion each, or a single stage gets slammed.  You'd probably need to breadboard and decide which sounds better (HINT: most commercial amps seem to distort a little at each stage, then knock the signal down to avoid clobbering any one gain stage).

[PRR]

> The traditional approach yielding the most success starts at the speaker/output section and works backwards towards the input jack.

Agreed.

However a thorough think-through will hike out-to-in and *also* in-to-out, making user-allowances along the way.

I'd want Full Roar to happen with 20mV in on the normal channel-- that determines total gain. Then I would apply 20mV to 500+mV input and see where it distorts. It should be possible to turn-down so-that >200mV input is not severely bent along the way until it just-clips the output. Or at another knob setting, 200mV input is not very strained internally when the power (co$t) stage is grossly overloaded (trying to make 2X to 10X the voltage it really can, output severely clipped).

Some larger amps also aim for gross clipping in some internal stage. That's the same as designing to clip the output stage except at some earlier point. With, obviously, another gain-knob after the clobbered stage so its sound can be reproduced cleanly to the speaker.

OTOH, plagiarism is IMHO much more common than anybody says. I'm looking at the early amps and their makers, and I'm not real sure they even had good voltmeters. I would doubt that Leo had a K&E slide-rule as one site speculates. (He did math, he worked as an accountant; but he was also "thrifty" and I'd guess an inexpensive used pearwood slip-stick, and he probably did more arithmetic on scrap-paper than slide-rule.)

[charczar27]

@charczar27:  You have asked a umber of intersting questions:  I suggest some reference books:
Dave Funk's Tube Amp Workbook
Inside Tube Amps, Dan Torres
Electric Guitar Amplifier Handboo, Jack Darr
TUT1, Kevin O'Connor (pretty technical)
The Tube Amp Book, Aspen Pittman

I will DEFINITELY look into these, thank you. A good book goes a long way.

> The traditional approach yielding the most success starts at the speaker/output section and works backwards towards the input jack.

Agreed.

However a thorough think-through will hike out-to-in and *also* in-to-out, making user-allowances along the way.

Yeah, since my original post I've went back and have been attempting to design using this methodology. It's been working a lot better than the in to out method.

I have also been running early Fender designs through Multisim and have noticed that nearly all of them are clipping early in the Preamp (i.e. before the PI). So I'm just going to push through and go with the design that I have and tweak from then on. Thank you guys for all the resources and help you've given!


**unrelated to original post**

However, I've come across an interesting phenomena that you guys may be able to answer....

In the Bassman and a few other circuits, I've noticed that a higher voltage is being run to the phase inverter. In most instances this voltage exceeds the maximum plate voltage listed on the tube's datasheet. For example, in the Bassman 5F6A the supply voltage is at 385 V and the actual plate voltage at 230 V. This build used a 12AX7 at the PI; max plate voltage of 12AX7 is 330 V. Also, in a later Bassman model a 12AT7 is used as the PI. Supply voltage is 420 V and the plate voltage is 340 V. Max plate voltage of the 12AT7 is 300 V. What gives? I thought I had nearly everything figured out but then I stumbled across this 

[jjasilli]

Leo  cheated in a number of ways.  The cathode follower voltage is technically wrong also.  Again, if you use google there's tons of info on the web, and reference books re the Bassman in its various incarnations.  E.g., The Fender Bassman 5F6-A, by Richard Kuehnel.  https://www.ampbooks.com/

[HotBluePlates]

... In most instances this voltage exceeds the maximum plate voltage listed on the tube's datasheet. ...

Plate-to-cathode voltage is what matters.

Even then, real-world amps sometimes exceeded the data sheet specs.  For example, the 12AX7 in V4 of the McIntosh MC-30 has 490v plate-to-cathode.

[PRR]

> a higher voltage is being run to the phase inverter.

Yes, but (as HBP says) the tube "never" has to eat ALL of that.

In 5F6a the tube plate-cathode voltage is like 200V, far shy of the 300V rating.

[charczar27]

Once again, thank you guys. I figured it out in a different way; the cathode resistor is 470, then the...is it called a tail resistor? Anyway, it's the 10K resistor - the same current is running through both resistors an while the 470 only produces a slight voltage (About 1 - 2 volts) the 10K is going to produce about 20x that amount. I've just verified on the schematic as well. I don't know how I never realized this. 32.5 V / 10K ohms = 3.25 mA; 3.25 mA * 470 Ohms = 1.5 V. 1.5 V + 32.5 V = 34 V. So anode to catode voltage in this case is 230 V - 34 V = 196 V.....Is this correct?

If so I made a big mistake in my LTP design 

[charczar27]

> a higher voltage is being run to the phase inverter.

Yes, but (as HBP says) the tube "never" has to eat ALL of that.

In 5F6a the tube plate-cathode voltage is like 200V, far shy of the 300V rating.

I just responded without reading this first.....question answered haha; thank you so much.